Wrapping images sensibly via VopSoP?

[oldiesgoodies]

Hi

I am having little hard hard time using VopSop to wrap an image around a sphere. When I use planar uv projection, all works great but when I change my texture to"polar" or "cylindrical", the colors on the sphere makes no sense at all.

Here is what I have

-created a poly sphere

-added uv texture

-added VopSop

-imported uv attrbs

-added color map

-fed the attrib first and second into color map`s uv slots

-and fed the output from colormap to Cd

As I mentioned this sceheme works when the projection is orthogonal onto the sphere, but every other projection method produces jibberish noisy color distribution.

I would appreciate if someone brings some insight into this

thanks

[anim]

it is probable that in UVtexture SOP you have mode set to Natural Location

this will create Point Attribute for Orthographic but for Polar and Cylindrical it is Vertex Attribute

and vertex attributes will not be read automatically by parameter VOP and it should produce nothing but single color

but if you are getting noisy color maybe try increasing polygon resolution

anyways try to change the mode to Point texture and see if it does anything

or append Attribute Promote and promote UV attribute from Vertex to Point

since you are creating point Cd anyways, vertex uvs are no use (even though they are more accurate)

if you want to use vertex uvs, use Facet SOP to make unique points, then promote Vertex uvs to Point, then your VOPSOP which will create point Cd then promote Point Cd to Vertex then fuse points

[oldiesgoodies]

anim, thanks for the insight. You are right about points vs vertices. When I try different options I get it to work. Why on earth different projection methods would make such structural difference? I never thought about it.

One thing I have also realized that this stuff also created confusion when I was working on some materials. Thanks for the help again.

Edited May 4, 2010 by oldiesgoodies

[anim]

...

When I try different options I get it to work. Why on earth different projection methods would make such structural difference?

...

because vertex uvs are more accurate and polar, cylindrical and some other projections need that accuracy otherwise they'll create nasty artifacts

(Fix Boundary Seams should be checked however)

Ortho projection will look the same as vertex or point so there is more sense to keep it as point attribute because it is much more memory/storage efficient

[oldiesgoodies]

anim, thanks for shedding light onto the issue.